Changes in Nucleosome Occupancy Associated with
Metabolic Alterations in Aged Mammalian Liver
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation
Bochkis, Irina M., Dariusz Przybylski, Jenny Chen, and Aviv
Regev. “Changes in Nucleosome Occupancy Associated with
Metabolic Alterations in Aged Mammalian Liver.” Cell Reports 9,
no. 3 (November 2014): 996–1006.
As Published
http://dx.doi.org/10.1016/j.celrep.2014.09.048
Publisher
Elsevier B.V.
Version
Final published version
Accessed
Thu May 26 03:20:35 EDT 2016
Citable Link
http://hdl.handle.net/1721.1/96715
Terms of Use
Creative Commons Attribution-NonCommercial-NoDerivs 3.0
Unported License
Detailed Terms
http://creativecommons.org/licenses/by-nc-nd/3.0/
Article
Changes in Nucleosome Occupancy Associated with
Metabolic Alterations in Aged Mammalian Liver
Graphical Abstract
Authors
Irina M. Bochkis, Dariusz Przybylski,
Jenny Chen, Aviv Regev
Correspondence
ibochkis@broadinstitute.org
In Brief
Bochkis et al. use genome-wide profiling
in the livers of young and old mice to
observe age-dependent changes in
PPARa targets as well as nucleosome occupancy linked to Foxa2 and Hdac3 binding sites. A reciprocal binding pattern of
Foxa2 and Hdac3 at PPARa targets contributes to gene-expression changes
that lead to steatosis in the aging liver.
Highlights
Accession Numbers
An in vivo genome-wide nucleosome map in the aging liver is
described
GSE58006
GSE57809
GSE58005
GSE60393
Foxa2 binds regions of decreased nucleosome occupancy at
PPARa targets in old livers
Hdac3 and Srf are implicated in age-dependent metabolic
dysfunction
Reciprocal binding of Foxa2 and Hdac3 at PPARa targets contributes to steatosis
Bochkis et al., 2014, Cell Reports 9, 996–1006
November 6, 2014 ª2014 The Authors
http://dx.doi.org/10.1016/j.celrep.2014.09.048
Cell Reports
Article
Changes in Nucleosome Occupancy
Associated with Metabolic Alterations
in Aged Mammalian Liver
Irina M. Bochkis,1,* Dariusz Przybylski,1 Jenny Chen,1,2 and Aviv Regev1,3
1Broad
Institute of MIT and Harvard, Cambridge, MA 02142, USA
of Health Sciences and Technology, Massachusetts Institute of Technology, Cambridge, MA 02142, USA
3Howard Hughes Medical Institute, Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
*Correspondence: ibochkis@broadinstitute.org
http://dx.doi.org/10.1016/j.celrep.2014.09.048
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
2Division
SUMMARY
Aging is accompanied by physiological impairments,
which, in insulin-responsive tissues, including the
liver, predispose individuals to metabolic disease.
However, the molecular mechanisms underlying
these changes remain largely unknown. Here, we
analyze genome-wide profiles of RNA and chromatin
organization in the liver of young (3 months) and old
(21 months) mice. Transcriptional changes suggest
that derepression of the nuclear receptors PPARa,
PPARg, and LXRa in aged mouse liver leads to
activation of targets regulating lipid synthesis and
storage, whereas age-dependent changes in nucleosome occupancy are associated with binding sites
for both known regulators (forkhead factors and nuclear receptors) and candidates associated with nuclear lamina (Hdac3 and Srf) implicated to govern
metabolic function of aging liver. Winged-helix transcription factor Foxa2 and nuclear receptor corepressor Hdac3 exhibit a reciprocal binding pattern
at PPARa targets contributing to gene expression
changes that lead to steatosis in aged liver.
INTRODUCTION
Aging is associated with increased prevalence of metabolic disease and cancer, reduced capacity for tissue regeneration, and
physical decline (Rodriguez et al., 2007; Willis-Martinez et al.,
2010). In particular, triglyceride accumulation is the common
metabolic phenotype of aging liver and of metabolic syndrome,
an age-related disorder that increases the risk of developing diabetes and cardiovascular disease. While mechanisms of agedependent defects in tissue regeneration have been studied
extensively (Jin et al., 2010; Jin et al., 2009), causes of age-onset
metabolic impairments remain largely unknown.
Previous studies suggest a role for chromatin and nuclear organization in aging-associated lipid dystrophies. First, changes
in chromatin organization mediate age-dependent impairments
in several tissues (Chambers et al., 2007; Jin et al., 2010).
Furthermore, mutations in lamin A/C (LMNA), a nuclear envelope protein, cause progeria, a premature aging syndrome,
and lamin-dependent defects have been connected to physiological human aging (Scaffidi and Misteli, 2006). LMNA is also
mutated in partial lipodystrophy (Shackleton et al., 2000), a condition associated with insulin-resistant diabetes and hepatic
steatosis. Clinical features of lipodystrophy due to mutations
in LMNA closely resemble those in individuals with mutations
in PPARG, a nuclear receptor involved in pathogenesis of fatty
liver (Gavrilova et al., 2003). Progeria attributed to defects in
DNA repair (Schumacher et al., 2009) is modeled in mice by
deletion of Ercc1, an enzyme crucial to nucleotide excision
repair (Niedernhofer et al., 2006). Ercc1 mutants exhibit upregulation of targets of the hepatic nuclear receptors (PPARa and
PPARg) and hepatic steatosis. Finally, liver-specific loss of
Foxa2, a pioneer transcription factor regulating nucleosome
dynamics, leads to premature aging, increased hepatic lipogenesis, and age-onset obesity (Bochkis et al., 2013). While the progeroid models described above propose a role for epigenetic
conformation in modulating age-dependent metabolic impairments, this hypothesis has not yet been tested in a model of
physiological aging.
To study the relation between chromatin organization and
age-dependent metabolic impairments in the liver, we examined
gene expression and genome-wide nucleosome occupancy in
livers isolated from young (3 months) and old (21 months) mice
(Figure 1A). Age-dependent induction of expression in lipid synthesis and storage genes, similar to metabolic changes seen in
progeroid syndromes, is consistent with derepression of the nuclear receptors PPARa, PPARg, and LXRa. Analysis of transcription factor binding sites that are overrepresented in regions
where nucleosome occupancy changes with age identified established regulators of age-dependent metabolic dysfunction
and lamina-associated candidates. Winged-helix transcription
factor Foxa2 that regulates nucleosome dynamics binds regions
of decreased nucleosome occupancy at PPARa targets in old
livers. Conversely, binding of nuclear receptor corepressor
Hdac3, detected from motifs found in regions of increased
nucleosome occupancy, exhibits a reciprocal pattern. Together,
altered Foxa2 and Hdac3 occupancy at PPARa targets contributes to gene expression changes that lead to steatosis in aged
liver.
996 Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors
A
C
D
B
Figure 1. Nuclear-Receptor-Dependent and Inflammatory Targets Are Induced during Physiological Aging in the Liver
(A) Experimental design includes comparing gene expression (RNA-seq) from livers of young (3 months), and old (21 months) and nucleosome occupancy
(MNase-seq) of young and old livers.
(B) RNA-seq track view in Integrative Genome Viewer (IGV) of expression of Cidea, a target of nuclear receptors Ppara and Pparg, not expressed in young healthy
livers and highly induced in steatotic older hepatocytes.
(C) Heatmap of differentially expressed genes in aged livers (FDR 5%, 727 upregulated and 525 downregulated).
(D) Regulators of gene expression in heat map in (C). Factors regulating lipid synthesis and storage are nuclear receptors Ppara (p value 1.1 3 10 20), Pparg
(p value 2.7 3 10 8), LXRa (p value 4.2 3 10 13), and Cebpb (p value 2.1 3 10 9). Additional regulators include NF-kB (p value 1.8 3 10 3), PI3K (p value 6.4 3
10 6), and Xbp1 (p value 3.6 3 10 4). Networks with upregulated and downregulated targets are shown in red and blue, respectively. p values were determined by
the Fisher’s exact test (using IPA).
RESULTS
Inflammatory and Nuclear Receptor Target Genes Are
Induced in Aged Liver
To characterize transcriptional changes in aging, we profiled
gene expression by RNA sequencing (RNA-seq) in young
(3 months) and old (21 months) male mice (three to four animals
per age group; Figure 1A) and identified 1,252 genes differentially expressed between young and old mice (727 induced
with age; 525 repressed; false discovery rate [FDR] 5% edgeR;
Robinson et al., 2010; Experimental Procedures; Table S2).
Consistent with previous studies of genetically induced aging
(Niedernhofer et al., 2006), genes induced with age were enriched for ‘‘Lipid & Fatty Acid Metabolism’’ functions (88 of
1,157 pathway genes, p value = 4.1 3 10 9, Fisher’s exact
test). Among the highly induced genes were Cidea, a target of
nuclear receptors PPARa and PPARg, encoding a lipid-associated protein only detected in fatty and diabetic livers (Gong
et al., 2009); related family members Cideb and Cidec, cytochrome p450 detoxification enzymes (Cyp2b9, Cyp2b10, and
Cyp2b13) involved in stress response; and histone H4 transcript
(Hist1h4c; Figures 1B–1D). In contrast, mRNA levels of Moxd1,
an enzyme in the endoplasmic reticulum (ER) and Asns, an
enzyme upregulated by ER stress response, are downregulated
in aged hepatocytes.
Genes induced with aging were also enriched for known
targets of key transcriptional regulators of lipid homeostasis,
including peroxisome proliferator activated receptors (PPARa
and PPARg; p values 1.1 3 10 20 and 2.7 3 10 8, respectively, Ingenuity Pathway Analysis [IPA], Fisher’s exact test),
liver X receptor (LXRa, Nr1h3, p value 4.2 3 10 13), PGC1a
(p value 1.8 3 10 4), a coactivator of PPARs, and CEBPB
(p value 2.1 3 10 9), known to coregulate PPARg targets
(Lefterova et al., 2008) (Figures 1D, S1A, and S1B; Table
S3). IPA also identifies gene networks regulated by agonists
for PPARa (pirinixic acid/WY-14643 and clofibrate, p values
9.6 3 10 20 and 1.8 3 10 7, respectively) and LXRa
(T0901317, p value 4.2 3 10 13), suggesting that PPARa
and LXRa are ligand activated in aged liver. In addition, genes
activated by inflammatory regulators (NF-kB/RELA, IRF3, and
Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors 997
Figure 2. Characterization of Regions of
Age-Dependent Change in Nucleosome
Occupancy
(A) MNase-seq tracks show nucleosome occupancy of young (3M, blue) and old (21M, purple)
biological replicates in a 2 kb region (top, unpaired;
bottom, paired).
(B) Examples of loss (full and partial, two top
panels) and gain (full and partial, two middle
panels) in nucleosome occupancy for paired (left)
and unpaired (right) replicates. Examples of loss
and gain in nucleosome occupancy overlapping a
CTCF binding site (bottom panel).
(C) The distribution of age-dependent nucleosome
occupancy change regions (both losses and gains)
around the TSS. Values for replicate 1 (in bold) and
replicate 2 are shown above each bar. The majority
of changes occur distally, 50 to 500 kb from the
TSS.
TLR4; p values 1.8 3 10 3, 2.9 3 10 2, and 2.1 3 10 3
respectively) are also upregulated in aged mice (Figures 1D,
S1A, and S1B; Table S3).
These results are consistent with a model where one or more
of these regulators is activated during aging, leading to
increased transcription of key lipid metabolism genes. Notably,
PPARs are upregulated in progeroid syndromes, LXRa is activated in prematurely aged Foxa2 mutants (Bochkis et al.,
2013), and inflammatory signaling is known to be activated by dietary fatty acids and contributes to insulin resistance (Fessler
et al., 2009).
Age-Dependent Changes in Nucleosome Occupancy
Are Enriched in Distal Elements
To study the role of chromatin organization in these age-dependent changes in transcription, we next measured genome-wide
nucleosome occupancy profiles in livers from young (3 months)
and old (21 months) mice. We used MNase digestion of chromatin followed by sequencing (MNase-seq; Experimental Proce-
dures; Umlauf et al., 2004), with two biological replicates for each age group:
replicate 1 with single-end sequencing,
and Replicate 2 with paired-end
sequencing (Figure 2A). We calculated
nucleosome occupancy for each sample
separately with the DANPOS package
(Chen et al., 2013), with moderately
good correlation between replicates
(Spearman r = 0.78 and 0.76 for young
and old replicates, respectively) and occupancy measures substantially higher
in the second replicate. Given the differences between the replicates (Figure S3A
and above), we calculated changes in
nucleosome occupancy between young
and old samples separately for each replicate and used DANPOS (Chen et al.,
2013) to identify regions with significant
occupancy changes of well-positioned nucleosomes (29,355
and 61,869 regions of loss and 24,979 and 61,540 regions of
gain for replicates 1 and 2, respectively; Figure 2B). Finally,
similar pathways are enriched for genes with changes in nucleosome occupancy in either replicate (Figure S3B).
Most age-dependent changes in nucleosome occupancy
occur distally, 50 to 500 kb from the transcription start site
(TSS; Figure 2C). Furthermore, there is an overlap between regions with changes in nucleosome occupancy and those bound
the insulator CTCF or marked with H3K4me1 (10% and 9% of
all regions with occupancy change for replicate 1 and 2,
respectively) or with H3K9ac (9% and 3% of all regions with occupancy change for replicate 1 and 2, respectively) in young
livers, modifications associated with enhancers and active
euchromatin, respectively. This suggests that distal enhancers,
known to be bound by Foxa factors and nuclear receptors
(Bochkis et al., 2012; Lefterova et al., 2008), may be the
genomic features most associated with nucleosome occupancy changes.
998 Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors
Foxa2 Binds Regions of Reduced Nucleosome
Occupancy at PPARa Targets in Aged Liver
To identify transcriptional regulators whose binding may be
affected by changes in nucleosome positions, we performed
both de novo motif discovery (using MEME; Bailey et al., 2009;
Experimental Procedures) and positional weight matrix (PWM)
scan analysis (using PscanChIP; Zambelli et al., 2013; Experimental Procedures) of 150 bp windows spanning all nucleosomes gained or lost with age for either replicate.
Motifs associated with forkhead transcription factors are the
most overrepresented in regions of changed occupancy (both
gain and loss) for both replicates. FOXO factors, transducers
of insulin signaling, can decompact chromatin (Hatta and Cirillo,
2007), and FOXA2 is known to bind nucleosomal DNA in vivo (Li
et al., 2011). Indeed, FOXA2 targets (defined as differentially expressed in 1-year-old Foxa2 mutants versus wild-type [WT] liver;
Bochkis et al., 2013) are particularly differentially expressed between our WT old versus young livers (Figure S3C; KolmogorovSmirnov [K-S] p value 2.4 3 10 9).
Motifs for nuclear receptors (RORa p value 3.1 3 10 56, LXR p
value 7.6 3 10 85, PPAR p value 9.7 3 10 76) and interferon regulatory factors (IRF p value 8.1 3 10 80) were highly enriched in
regions of nucleosome occupancy loss (Figure 3A). Protein
expression of the lipid metabolism regulators FOXA2 and PPARa
is not altered in aged livers, while that of PPARg is induced,
consistent with previous reports of PPARg induction in fatty liver
(Panasyuk et al., 2012) (Figure 3B). IPA of genes with nucleosome occupancy loss near TSSs identified networks regulated
by ligand-activated nuclear receptors PPARa, PPARg, and
LXRa. Expression was induced for one-quarter of these targets
(fold change R 1.3; Figure 3C).
Since Foxa2 plays an important role in lipid metabolism in
aged liver (Bochkis et al., 2013), we next examined Foxa2 binding in young and old livers using chromatin immunoprecipitation
sequencing (ChIP-seq). Strikingly, Foxa2 occupies substantially
more sites in old hepatocytes (12,834) than in young ones (6,605;
Figure 3D). Additional bound regions in the old livers are found at
PPAR target genes as well as at the Ppara promoter (Figure 3E).
The increase in binding is not due to change in FOXA2 expression (Figure 3B).
A subset of 734 Foxa2 sites that are ‘‘gained’’ in old livers
correspond to regions of decreased nucleosome occupancy.
Interestingly, while the majority of changes in nucleosome occupancy occur distally, these sites are found mostly at promoters
(Figure 3F). Genes associated with the binding sites are enriched
in functional categories including ‘‘hepatic steatosis’’ (p value
1.5 3 10 4), ‘‘nuclear import’’ (p value 3.8 3 10 4), and
‘‘increased circulating VLDL’’ (p value 3.5 3 10 5). In addition,
the PPAR/DR-1 element is enriched at these sites, suggesting
that age-dependent Foxa2 binding could enable PPAR binding
at these regions. Alternatively, newly bound FOXA2 at the promoter may interact with PPAR proteins that are bound to existing
distal enhancer elements to activate PPAR targets (Figure 6).
Increased Nucleosome Occupancy Suggests a Role for
cKrox-Hdac3 in Age-Dependent Dysfunction
Regions associated with increased nucleosome occupancy
were enriched, among other elements, with a motif bound by
Srf (p value 3.3 3 10 45), a factor that interacts with the nuclear
lamina (Swift et al., 2013) (Figure 3A) and with motifs identified as
a GAGA repeat (short motif p value 1.8 3 10 111, long motif p
value 2.3 3 10 8; Figure 4A). The short GAGA motif is frequent
along the genome (background frequency 0.416), but the prevalence of this sequence in the regions of age-dependent occupancy gain was higher (frequency 0.527). The GAGA repeat is
associated with lamina-associated domains (LADs) (Lund
et al., 2013) and is bound by the transcriptional repressor cKrox
(Zbtb7b) in a complex with Hdac3. While SRF protein levels
decrease, expression of HDAC3 is not altered in older livers (Figure 4B). Hdac3 can act as a corepressor of nuclear receptors
regulating hepatic lipid metabolism genes (Knutson et al.,
2008; Sun et al., 2012). Furthermore, the Ncor complex, an activating cofactor for Hdac3 (Guenther et al., 2001), is predicted to
be repressed in our aged livers based on the expression of its
known targets (IPA; Experimental Procedures). Taken together,
this suggests that Hdac3 activity may also be reduced in older
animals, resulting in induction of its targets by nuclear receptors.
To test this hypothesis, we first compared the genes differentially expressed between (young) Hdac3 liver-specific mutants
versus WT (Knutson et al., 2008) with the genes differentially expressed between aged and young livers. We found that the
Hdac3-dependent genes are particularly differentially expressed
in old versus young livers (compared to the background of all old
versus young differentially expressed genes; K-S p value 2.2 3
10 16). In addition, genes from the same biological functions
and pathways were affected in both young Hdac3 mutant
mice and differentially expressed genes in aged livers (Figures
4C and 4D).
Next, we used ChIP-seq to measure Hdac3 binding in young
and old livers and found markedly less occupied regions in old
livers (5,828 in young, 2,930 in old; Figure 4E), especially at
PPAR-dependent genes. PPAR response element (DR-1) was
significantly overrepresented in the sequences bound by
Hdac3 in young, but not old, livers (p value 1.4 3 10 31). While
the overlap between Hdac3 sites bound in young livers and regions of increased nucleosome occupancy is small (152 regions),
the GAGA motif is also enriched at sites bound by Hdac3 in the
young livers (short motif p value 7.0 3 10 44, long motif p value
3.6 3 10 13; Figure 4F), suggesting the overlap could be greater
with deeper coverage of MNase-seq data and more statistically
significant nucleosome occupancy changes calls.
Sites for serum response factor (SRF) were also enriched in regions with age-dependent gain of nucleosome occupancy (p
value < 6 3 10 74, PScan ChIP analysis for overrepresented
PWMs). Furthermore, SRF-dependent genes were differentially
expressed in older livers (IPA analysis; Experimental Procedures;
Figure 4G), similar gene functions are enriched in genes differentially expressed in liver-specific Srf mutants (versus WT) or in old
liver (versus young; Figure 4H), and most (36 out of 48) Srf target
genes with age-dependent gain in nucleosome occupancy near
the TSS are induced in aged livers (Figure 4I). Altogether, these
data suggest a downregulation of Srf activity with aging, consistent with reduced SRF protein levels (Figure 4B), leading to derepression of its gene targets. Srf activity has also been shown to
be repressed during senescence by nuclear exclusion (Ding
et al., 2001).
Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors 999
Figure 3. Foxa2 Binds PPARa Targets in Old Liver
(A) PWM scan analysis identifies numerous forkhead PWMs and nuclear receptor motifs (Rora p value 3.1 3 10 56, LXR p value 7.6 3 10 85, PPAR p value 9.7 3
10 76) in regions of age-dependent loss of nucleosome occupancy. PPAR factors bind a direct repeat (DR-1 element). The repeats are enclosed by black
rectangles. Forkhead PWMs, as well as matrices for Srf (p value 3.3 3 10 45) and Rreb1 (p value 5.4 3 10 98), and a sequence resembling a telomeric repeat
bound by Rap1 (p value 5.9 3 10 64), are significantly overrepresented in regions of gain of nucleosome occupancy.
(B) Western blot analysis of protein nuclear extracts from four young (3 months) and four old (21 months) mouse livers with antibodies to FOXA2, PPARa, PPARg,
and TATA box binding protein (TBP, loading control).
(C) Heatmaps showing expression of Ppara and Pparg targets (left) and LXRa targets (right) with age-dependent nucleosome occupancy loss near the TSS.
(D) Venn diagram showing the results of genome-wide location analysis for Foxa2 in young and old liver, identifying 6,605 binding sites in young and 12,834 in old,
of which 2,738 were called bound by both factors by PeakSeq.
(E) ChIP-seq track view in IGV of increased Foxa2 binding in old liver at the Ppara locus.
(F) The distribution of regions of additional Foxa2 binding in old livers that correspond to loss of nucleosome occupancy around the TSS. While the majority of
changes in nucleosome occupancy occur distally, these sites are found mostly at promoters.
Foxa2 and Hdac3 Exhibit Reciprocal Binding Pattern
at PPARa Targets
Strikingly, we found that Foxa2, a pioneer factor that binds the
forkhead motif, occupies substantially more regions in old livers
(6,605 in young, 12,834 in old), whereas lamina-associated nuclear receptor corepressor Hdac3 binds markedly less regions
(5,828 in young, 2,930 in old). These differential binding sites
for both factors are found at functional PPARa target genes
(Rakhshandehroo et al., 2010) (Figure 5) and contain the PPAR
response element. This is consistent with a model where
Foxa2 binding likely leads to nucleosome eviction in older livers.
Foxa2 cooperates with PPAR receptors, either enabling PPAR
1000 Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors
binding at the promoter or interacting with existing PPAR
proteins bound to enhancer elements, leading to upregulation
of targets involved in lipid synthesis and storage (Figure 6A).
Hdac3 may regulate hepatic lipid targets in either of two ways:
(1) through GAGA sites bound by cKrox/Hdac3 or (2) by repressing PPAR sites in young, but not old, livers (Figure 6B).
Together, the reciprocal binding pattern of Foxa2 and Hdac3
contributes to gene expression changes leading to steatosis in
aged liver.
DISCUSSION
Here, we used an unbiased approach to find candidate regulators that affect age-dependent metabolic dysfunction. Since nucleosomes and transcription factors compete for DNA binding
(Workman and Kingston, 1992), mapping genome-wide nucleosome composition and tracking changes in nucleosome occupancy in aged mice in vivo allowed us to test for differences in
transcription factor binding that are responsible for downstream
gene regulation governing age-dependent phenotypes.
Motifs bound by forkhead transcription factors and nuclear receptors are significantly overrepresented in regions of agedependent loss of nucleosome occupancy. We have examined
binding of Foxa2 in young and old livers, and it is likely that other
Fox factors, especially Foxa1 and Foxa3 and members of the
Foxo subfamily, could play a role in this process, and that possibility should be explored further. While nucleosome occupancy
dynamics observed in aged livers associates with distal enhancers, elements bound by forkhead transcription factors and
nuclear receptors in young livers (Bochkis et al., 2012; Lefterova
et al., 2008), we find that most Foxa2 sites that are bound only in
old livers and correspond to regions of decreased nucleosome
occupancy are found near the promoters. These sites are also
enriched for the PPAR/DR-1 element, suggesting that additional
Foxa2 binding might enhance accessibility and enable recruitment of PPAR factors to these elements (Figure 6A). We also
observe upregulation of PPAR-dependent gene expression for
genes with a nucleosome loss at the promoter. A recent study
has challenged the classical model of nuclear-receptor-dependent gene regulation, reporting that LXRa and PPARa binding
to their target loci in the liver is largely ligand dependent, with
the agonists enabling the receptors to occupy less accessible
sites (Boergesen et al., 2012). Two additional reports involving
progesterone receptor and estrogen receptor showed that
nucleosome occupancy observed in unstimulated cells is significantly depleted upon hormone activation (Ballaré et al., 2013;
Tropberger et al., 2013), allowing for nuclear receptor binding.
Our findings are consistent with this revised model and suggest
that nucleosome dynamics may mediate ligand-dependent activation of ‘‘metabolic’’ nuclear receptors.
While Foxa2 binding sites are also enriched for the PPAR/
DR-1 element, we cannot pinpoint which PPAR receptor
(PPARa, PPARg, or PPARd) binds these sites and in which
physiological condition. PPARa mediates the hepatic fasting
response, and binding of this factor should also be examined
in the fasted state. Hence, binding of PPAR receptors should
be explored in young and old livers to determine the relationship
between the factors and their roles in aged livers.
We find that shifts in hepatic gene expression in physiological
aging mirror differences observed in progeroid conditions.
Changes in nucleosome occupancy are associated with our inferred derepression of nuclear receptors regulating hepatic lipid
metabolism, leading to fatty liver (Figure 6). Examining changes
in nucleosome occupancy in vivo highlighted lamina-associated
regulators, Hdac3 and Srf, whose role in age-dependent metabolic dysfunction should be explored further.
Histone deacetylases related to Hdac3, Hdac1, and Sirt1 are
known to play important roles in aging liver (Jin et al., 2011; Willis-Martinez et al., 2010). Liver-specific deletion of Hdac3 leads
to fatty liver, a phenotype associated with aging, due to derepression of nuclear hormone receptor-dependent gene expression (Sun et al., 2012; Knutson et al., 2008). Hdac3 mutant livers
also exhibit upregulation of mTOR signaling similar to a model of
premature aging due to hepatocyte-specific ablation of Foxa2
(Bochkis et al., 2013). Deletion of Hdac3 also impacts DNA repair
and reduces heterochromatin content, as observed in aging
nuclei (Bhaskara et al., 2010). Loss of Hdac3 binding and transcriptional derepression of targets is observed in adipocytes in
a mouse model of progeria (Karakasilioti et al., 2013). Hence, it
is likely that Hdac3 is a pivotal regulator of epigenetic and metabolic changes during chronological aging.
The second candidate, Srf, regulates liver proliferation, hepatic lipid metabolism, and growth hormone/Igf-1 signaling crucial
to longevity (Sun et al., 2009). Transcription factors, including
Hif1a, Hsf1, and Xbp1, that govern different stress responses,
similar to Srf, affect gene expression during aging (Henis-Korenblit et al., 2010; Hsu et al., 2003; Kang et al., 2005). Loss of
Srf in the liver also alters mRNA levels of histone proteins and
chromatin regulators, similar to changes seen in aged livers. A
recent study reported that lamin A regulates Srf mRNA levels
and Srf-dependent gene transcription (Swift et al., 2013),
providing another link to aging.
Notably, ‘‘nuclear lumen’’ genes, including a number of histone transcripts, were highly overrepresented in targets changed
in older livers. Histone expression has been reported to decline in
a number of aging paradigms (Feser et al., 2010; Celona et al.,
2011; Liu et al., 2013). In contrast, we found that whereas
some histone transcripts are downregulated with age, others
are upregulated (Figures S2A–S2C). Downregulated histone H2
transcripts included replication-dependent (Hist2h2aa and
Hist1h2b) and replication-independent genes (H2afx). H2afx is
the principal chromatin component involved in DNA repair, and
reduced levels of this histone could explain defects in DNA repair
in aged livers. Histone variants differ in stability and DNA binding
and play distinct functions in the nucleus (Talbert and Henikoff,
2010). Changing composition of histone variants in aged tissues
in vivo could impact gene regulation and should be investigated
further.
Premature aging, due to either mutation in lamin A or defects in
DNA repair, is associated with dysregulation of lipid homeostasis
and upregulation of PPAR-dependent gene expression (Niedernhofer et al., 2006; Savage, 2009). We find that similar
pathways, also implicated in metabolic syndrome, are perturbed
in chronologically aged livers. We suggest a relationship between lamina-associated factors and age-dependent dysregulation of hepatic lipid metabolism. Whether lamina-dependent
Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors 1001
Figure 4. Regulators Associated with Nuclear Lamina Are Implicated in Mediating Age-Dependent Dysfunction in the Liver
(A) GAGA repeat motif (short motif p value 1.8 3 10 111, long motif p value 2.3 3 10 8), bound by a transcriptional repressor cKrox (Zbtb7b) at the nuclear lamina,
and an A/T rich sequence associated with lamina-associated domains (LADs), are enriched in regions of gain of nucleosome occupancy.
(B) Western blot analysis of protein nuclear extracts from four young (3 months) and four old (21 months) mouse livers with antibodies to SRF, HDAC3, and TATA
box binding protein (TBP, loading control).
(C and D) Comparison of overrepresented biological functions (C) and pathways (D) between differentially expressed genes in aged WT livers and young liverspecific Hdac3 mutants (K-S p value 2.2 3 10 16).
(E) Venn diagram showing the results of genome-wide location analysis for Hdac3 in young and old liver, identifying 5,828 binding sites in young and 2,930 in old,
of which 168 were called bound by both factors by PeakSeq.
(F) GAGA repeat motif (short motif p value 7.0 3 10 44, long motif p value 3.6 3 10 13) is also enriched in regions bound by Hdac3 in young liver.
(legend continued on next page)
1002 Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors
Figure 5. Foxa2 and Hdac3 Exhibit Reciprocal Binding Pattern at PPARa Targets
(A and B) Chip-seq track view of (A) Foxa2 and (B)
Hdac3 binding in young and old livers at the loci
encoding Ppara targets.
mechanisms could mediate age-onset degeneration in other tissues remains to be explored.
EXPERIMENTAL PROCEDURES
Young and Old Mice
Male mice (C57BL6) were purchased from the National Institute of Aging aged
rodent colony (Charles River Laboratories). Four young (3 months), and three
old (21 months) mice were used for RNA studies. Two biological replicates
of young and old mice were used for microccocal nuclease digestion and
sequencing. Two biological replicates of young and old mice were used for
chromatin immunoprecipitation and sequencing. Animals were housed for a
week after arrival to acclimate to the light-dark cycle at the Massachusetts
Institute of Technology (MIT) facility before tissue harvest. All animal work
was approved by MIT’s Committee on Animal Care (CAC protocol number
Regev-0612-058-15).
RNA Isolation and Profiling
Liver RNA isolation and quantitative real-time PCR was performed as previously described (Bochkis et al., 2013). Total RNA (2 mg) was enriched for
mRNA (Dynabeads mRNA direct kit, Life Technologies) used for strand-specific library preparation. mRNA was fragmented (Ambion fragmentation buffer,
70 C for 3 min), dephosphorylated, and concentrated (RNA Clean & Concentrator columns, Zymo Research). Ligation with RNA adapters was followed by
first-strand cDNA synthesis (AffinityScript reverse transcriptase enzyme). RNA
was degraded (EDTA and NaOH) while adapters containing barcodes were
ligated to cDNA, which subsequently was amplified for ten cycles of PCR
(KAPA HiFi DNA Polymerase, Kapa Biosystems). Libraries were sequenced
on an Illumina HiSeq 2500 (31 bp paired-end reads; Table S1).
Microccocal Nuclease Digestion and Sequencing
Digestion was performed on snap-frozen liver, as previously described (Umlauf et al., 2004) with a few modifications. Briefly, 100 mg of tissue was pulverized (Covaris cryoPrep impactor), and nuclei were purified using a
sucrose gradient. The nuclei were digested with MNase (Roche catalog number 10107921001) at 37 C for 12 min. Several titrations of MNase enzyme
were tested. Samples that were well digested, with most of the input as a
mononucleosome (150 bp) and a slight band of a dinucleosome (300 bp)
visible on a BioAnalyzer trace, to prevent overdigestion, were used for subsequent library preparation. Libraries (starting material 10 ng) were made as
previously described (Bochkis et al., 2012) and sequenced on an Illumina
HiSeq 2000 instrument (36 bp single-end and 25 bp paired-end read; summary in Table S4).
ChIP and ChIP-Seq
Snap-frozen mouse liver (100 mg) from WT mice
was used to prepare chromatin. ChIP and ChIPseq were performed as reported previously (Bochkis et al., 2012). A slight modification involved using
multiplex adapters for sequencing and Kapa HiFi
DNA polymerase (Kapa Biosystems) for PCR
amplification. FOXA2-specific rabbit antiserum
(Seven Hills Bioreagents, WRAB-1200) and rabbit antibody to HDAC3 (Abcam,
ab7030) were used for immunoprecipitation. Libraries were sequenced on an
Illumina HiSeq 2500 with 60 bp single-end reads (Table S6).
Western Blot Analysis
Nuclear extracts preparation and protein immunoblot analysis were performed
as reported previously (Bochkis et al., 2008). The primary antibodies used were
rabbit antibody to FOXA2 (Seven Hills Bioreagents, WRAB-1200, 1:5,000),
rabbit antibody to HDAC3 (Abcam, ab7030, 1:5,000), rabbit antibody to
PPARa (Santa Cruz, sc-9000, 1:100), rabbit antibody to PPARg (Santa Cruz,
sc-7196, 1:100), rabbit antibody to SRF (Santa Cruz, sc-335, 1:200), and rabbit
antibody to TBP (Santa Cruz, sc-273, 1:100).
RNA-Seq Data Analysis
RNA-seq reads were aligned with TopHat (Trapnell et al., 2009) to mouse
genome build mm9 (library and alignment statistics in Table S1). Expression
levels were calculated using RSEM (Li and Dewey, 2011). Differential expression analysis of RNA-seq (p value < 0.05) was performed in R using EdgeR
package (Robinson et al., 2010), with a Benjamini-Hochberg FDR of 5%.
Functional Analysis
Ingenuity curates information gathered from the literature, as well as genomic
experiments (microarray, RNA-seq, and ChIP-seq) to determine sets of targets
controlled by a regulator (a transcription factor, chromatin remodeler, kinase,
small molecule, etc.) and the logic of this regulation (activation or repression).
IPA’s analysis method compares the overlap of each such set of targets with
the subset of genes that are controlled by each regulator in an experimental
gene list and reports the p value of the overlap (Fisher’s exact test). If the overlap is significant, and if the expression changes agree with those expected
from the regulatory connection (genes activated by the regulator are activated
in the analysis data set and repressed genes are repressed, for example), IPA
analysis predicts whether the regulator associated with the gene targets is itself activated or inhibited. Analysis of overrepresented functional categories
and upstream regulators in IPA and heatmap generation was performed as
described previously (Bochkis et al., 2013).
ChIP-Seq Analysis
Reads were aligned to the mouse genome (mm9; NCBI Build 37; Table S6) using BWA (Li and Durbin, 2009). Duplicate reads were removed. Reads (phred
score > 20) that aligned uniquely were used for subsequent analysis. Data from
two biological replicates were merged for each condition (Foxa2 young, Foxa2
old, Hdac3 young, and Hdac3 old). PeakSeq (Rozowsky et al., 2009) was used
to identify bound peaks against input controls (Foxa2: FDR 10%, q-value =
0.12; Hdac3: FDR 5%, q-value = 0.2).
(G) IPA Analysis of Srf-dependent network of genes differentially expressed in older livers suggests that Srf activity is downregulated with aging (blue line
represents activation, orange line repression, and gray line the association with expression change).
(H) Comparison of overrepresented biological functions in aged WT-type livers and young liver-specific Srf mutants.
(I) Heatmap of Srf target genes with age-dependent gain in nucleosome occupancy (replicate 2; pattern similar for replicate 1).
Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors 1003
Figure 6. A Model Relating Chromatin
Changes to Development of Fatty Liver during Aging
Change in nucleosome occupancy in aged liver is
associated with upregulation of nuclear receptor
targets and development of steatosis.
(A) Foxa2 binding leads to nucleosome eviction in
older livers. Foxa2 cooperates with ligand-activated PPAR receptors (L, ligand), either interacting
with existing PPAR proteins bound to enhancer
elements (left) or enabling additional PPAR binding
at the promoter (right), leading to upregulation of
targets regulating lipid synthesis and storage.
(B) Hdac3 regulates hepatic lipid targets in two
ways: (1) at the nuclear lamina through GAGA sites
bound by cKrox/Hdac3 (left), and (2) by repressing
PPAR sites in young, but not old, livers in a classical
mechanism of nuclear receptor (NR) action. Regions with a GAGA motif, bound by cKrox (Zbtb7b)
in complex with Hdac3, inhibit expression of lipogenic targets in young livers. Age-dependent gain in nucleosome occupancy at these locations leads to eviction
of histone deacetylase Hdac3. Nucleosomes can now be acetylated, leading to active transcription of nuclear receptor targets. In addition, in a classical model of
nuclear receptor activation, Hdac3 binds unliganded PPARa in young livers and is evicted upon agonist stimulation in old livers. A coactivator with histone
acetyltransferase activity is recruited to the PPAR complex. Nearby nucleosomes are acetylated, and gene expression of the targets is turned on. The reciprocal
binding pattern of Foxa2 and Hdac3 at loci encoding PPARa targets contributes to dysregulation of hepatic lipid homeostasis during aging.
MNase-Seq Analysis
Reads were aligned using BWA (Li and Durbin, 2009) to mm9. Duplicate reads
(replicate 1) and fragments (replicate 2) were removed from the analysis. Reads
(replicate 1) and fragments (replicate 2; phred score > 30) that aligned uniquely
were used for subsequent analysis. Reads for each sample were down
sampled to the same number (150 million). Nucleosome calls (8.5 million
nucleosome peaks) and changes in nucleosome occupancy were calculated
by the DANPOS software (p value < 1 3 10 7 in Poisson test, parameters:
w = 150, d = 150, extend = 148 for single end, 74 for paired end).
chi-square test was used to assess whether a given chromatin mark is overrepresented in the set of age-dependent nucleosome occupancy changes
(as compared to a background set).
Regions Annotation
Regions of changed nucleosome occupancy were associated to closest genes
with the GREAT analysis tool (McLean et al., 2010). Overlap with histone modifications and CTCF binding (ENCODE data, 8-week mouse liver (Karolchik
et al., 2014) was computed using Galaxy genome analysis tools (Hillman-Jackson et al., 2012). Sequencing reads were visualized with the Integrative
Genome Viewer (IGV) (Robinson et al., 2011). To compute correlations between the biological replicates, we trimmed reads from replicate one (single
end) to 25 bp and disregarded pairing information in reads from replicate
two (paired end) and used single reads (read 1) of 25 bp. Reads were aligned
using BWA to Mus musculus genome build mm9, and duplicate reads were
removed from the analysis. Each read alignment was extended to 150 bp
(size of a nucleosome), and resulting data were used to compute genome
coverage and Spearman correlation.
SUPPLEMENTAL INFORMATION
Motif Analysis
MEME was used for de novo motif finding (default parameters: zero or one
motif per sequence, minimum width = 6, maximum width = 50, maximum number of motifs = 3) (Bailey et al., 2009). PscanChIP (Zambelli et al., 2013) was
utilized to identify overrepresented PWMs from Jaspar Version 5.0_Alpha
(Vlieghe et al., 2006) and Transfac 7.0 Public (Matys et al., 2006) databases
(parameters: assembly = mm9, background = mixed) among sequences corresponding to age-dependent change in nucleosome occupancy (150 bp
nucleosome-wide regions with significant change in occupancy as calculated
by DANPOS).
Statistical Analysis
The Kolmogorov-Smirnov test was performed to test the distributions of fold
changes of targets differentially expressed in both mutant mice (Foxa2 [Bochkis et al., 2013] and Hdac3 [Knutson et al., 2008]) and in aged WT livers and fold
changes of all genes differentially expressed in old versus young livers. The
ACCESSION NUMBERS
Genomic data from this study have been deposited to the NCBI Gene Expression Omnibus under accession number GSE58006 (GSE57809 for RNA-seq,
GSE58005 for MNase-seq, and GSE60393 for ChIP-seq).
Supplemental Information includes three figures and six tables and can be
found with this article online at http://dx.doi.org/10.1016/j.celrep.2014.09.048.
AUTHOR CONTRIBUTIONS
I.M.B. developed the project, performed experiments and data analysis, and
wrote the manuscript. D.P. and J.C. analyzed the data. A.R. edited the manuscript and provided funding for the study.
ACKNOWLEDGMENTS
We thank D. Thompson and R. Majovsky for critical reading of the manuscript
and L. Gaffney, R. Raychowdhury, H. Whitton, and I. Wortman for technical
assistance. This work was supported by NHGRI CEGS P50 HG006193 and
HHMI (A.R.). I.M.B. was supported by National Diabetes and Digestive and
Kidney Diseases Institute K01 award DK-101633.
Received: May 16, 2014
Revised: August 15, 2014
Accepted: September 24, 2014
Published: October 23, 2014
REFERENCES
Bailey, T.L., Boden, M., Buske, F.A., Frith, M., Grant, C.E., Clementi, L., Ren,
J., Li, W.W., and Noble, W.S. (2009). MEME SUITE: tools for motif discovery
and searching. Nucleic Acids Res. 37 (Web Server issue), W202–8.
Ballaré, C., Castellano, G., Gaveglia, L., Althammer, S., González-Vallinas, J.,
Eyras, E., Le Dily, F., Zaurin, R., Soronellas, D., Vicent, G.P., and Beato, M.
1004 Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors
(2013). Nucleosome-driven transcription factor binding and gene regulation.
Mol. Cell 49, 67–79.
Bhaskara, S., Knutson, S.K., Jiang, G., Chandrasekharan, M.B., Wilson, A.J.,
Zheng, S., Yenamandra, A., Locke, K., Yuan, J.L., Bonine-Summers, A.R.,
et al. (2010). Hdac3 is essential for the maintenance of chromatin structure
and genome stability. Cancer Cell 18, 436–447.
Bochkis, I.M., Rubins, N.E., White, P., Furth, E.E., Friedman, J.R., and Kaestner, K.H. (2008). Hepatocyte-specific ablation of Foxa2 alters bile acid homeostasis and results in endoplasmic reticulum stress. Nat. Med. 14, 828–836.
Bochkis, I.M., Schug, J., Ye, D.Z., Kurinna, S., Stratton, S.A., Barton, M.C., and
Kaestner, K.H. (2012). Genome-wide location analysis reveals distinct transcriptional circuitry by paralogous regulators Foxa1 and Foxa2. PLoS Genet.
8, e1002770.
Bochkis, I.M., Shin, S., and Kaestner, K.H. (2013). Bile acid-induced inflammatory signaling in mice lacking Foxa2 in the liver leads to activation of mTOR and
age-onset obesity. Mol Metab 2, 447–456.
Boergesen, M., Pedersen, T.A., Gross, B., van Heeringen, S.J., Hagenbeek,
D., Bindesbøll, C., Caron, S., Lalloyer, F., Steffensen, K.R., Nebb, H.I., et al.
(2012). Genome-wide profiling of liver X receptor, retinoid X receptor, and
peroxisome proliferator-activated receptor a in mouse liver reveals extensive
sharing of binding sites. Mol. Cell. Biol. 32, 852–867.
Celona, B., Weiner, A., Di Felice, F., Mancuso, F.M., Cesarini, E., Rossi, R.L.,
Gregory, L., Baban, D., Rossetti, G., Grianti, P., et al. (2011). Substantial histone reduction modulates genomewide nucleosomal occupancy and global
transcriptional output. PLoS Biol. 9, e1001086.
Chambers, S.M., Shaw, C.A., Gatza, C., Fisk, C.J., Donehower, L.A., and
Goodell, M.A. (2007). Aging hematopoietic stem cells decline in function and
exhibit epigenetic dysregulation. PLoS Biol. 5, e201.
Chen, K., Xi, Y., Pan, X., Li, Z., Kaestner, K., Tyler, J., Dent, S., He, X., and Li, W.
(2013). DANPOS: dynamic analysis of nucleosome position and occupancy by
sequencing. Genome Res. 23, 341–351.
Ding, W., Gao, S., and Scott, R.E. (2001). Senescence represses the nuclear
localization of the serum response factor and differentiation regulates its nuclear localization with lineage specificity. J. Cell Sci. 114, 1011–1018.
Feser, J., Truong, D., Das, C., Carson, J.J., Kieft, J., Harkness, T., and Tyler,
J.K. (2010). Elevated histone expression promotes life span extension. Mol.
Cell 39, 724–735.
Fessler, M.B., Rudel, L.L., and Brown, J.M. (2009). Toll-like receptor signaling
links dietary fatty acids to the metabolic syndrome. Curr. Opin. Lipidol. 20,
379–385.
Gavrilova, O., Haluzik, M., Matsusue, K., Cutson, J.J., Johnson, L., Dietz, K.R.,
Nicol, C.J., Vinson, C., Gonzalez, F.J., and Reitman, M.L. (2003). Liver peroxisome proliferator-activated receptor gamma contributes to hepatic steatosis,
triglyceride clearance, and regulation of body fat mass. J. Biol. Chem. 278,
34268–34276.
Gong, J., Sun, Z., and Li, P. (2009). CIDE proteins and metabolic disorders.
Curr. Opin. Lipidol. 20, 121–126.
Guenther, M.G., Barak, O., and Lazar, M.A. (2001). The SMRT and N-CoR corepressors are activating cofactors for histone deacetylase 3. Mol. Cell. Biol.
21, 6091–6101.
Hatta, M., and Cirillo, L.A. (2007). Chromatin opening and stable perturbation
of core histone:DNA contacts by FoxO1. J. Biol. Chem. 282, 35583–35593.
Henis-Korenblit, S., Zhang, P., Hansen, M., McCormick, M., Lee, S.J., Cary,
M., and Kenyon, C. (2010). Insulin/IGF-1 signaling mutants reprogram ER
stress response regulators to promote longevity. Proc. Natl. Acad. Sci. USA
107, 9730–9735.
Hillman-Jackson, J., Clements, D., Blankenberg, D., Taylor, J., and Nekrutenko, A.; Galaxy Team (2012). Using Galaxy to perform large-scale interactive
data analyses. Curr. Protoc. Bioinformatics Chapter 10, 5.
Hsu, A.L., Murphy, C.T., and Kenyon, C. (2003). Regulation of aging and agerelated disease by DAF-16 and heat-shock factor. Science 300, 1142–1145.
Jin, J., Wang, G.L., Timchenko, L., and Timchenko, N.A. (2009). GSK3beta and
aging liver. Aging (Albany, N.Y. Online) 1, 582–585.
Jin, J., Wang, G.L., Iakova, P., Shi, X., Haefliger, S., Finegold, M., and Timchenko, N.A. (2010). Epigenetic changes play critical role in age-associated
dysfunctions of the liver. Aging Cell 9, 895–910.
Jin, J., Iakova, P., Jiang, Y., Medrano, E.E., and Timchenko, N.A. (2011). The
reduction of SIRT1 in livers of old mice leads to impaired body homeostasis
and to inhibition of liver proliferation. Hepatology 54, 989–998.
Kang, M.J., Kim, H.J., Kim, H.K., Lee, J.Y., Kim, D.H., Jung, K.J., Kim, K.W.,
Baik, H.S., Yoo, M.A., Yu, B.P., and Chung, H.Y. (2005). The effect of age
and calorie restriction on HIF-1-responsive genes in aged liver. Biogerontology
6, 27–37.
Karakasilioti, I., Kamileri, I., Chatzinikolaou, G., Kosteas, T., Vergadi, E., Robinson, A.R., Tsamardinos, I., Rozgaja, T.A., Siakouli, S., Tsatsanis, C., et al.
(2013). DNA damage triggers a chronic autoinflammatory response, leading
to fat depletion in NER progeria. Cell Metab. 18, 403–415.
Karolchik, D., Barber, G.P., Casper, J., Clawson, H., Cline, M.S., Diekhans, M.,
Dreszer, T.R., Fujita, P.A., Guruvadoo, L., Haeussler, M., et al. (2014). The
UCSC Genome Browser database: 2014 update. Nucleic Acids Res. 42 (Database issue), D764–D770.
Knutson, S.K., Chyla, B.J., Amann, J.M., Bhaskara, S., Huppert, S.S., and Hiebert, S.W. (2008). Liver-specific deletion of histone deacetylase 3 disrupts
metabolic transcriptional networks. EMBO J. 27, 1017–1028.
Lefterova, M.I., Zhang, Y., Steger, D.J., Schupp, M., Schug, J., Cristancho, A.,
Feng, D., Zhuo, D., Stoeckert, C.J., Jr., Liu, X.S., and Lazar, M.A. (2008).
PPARgamma and C/EBP factors orchestrate adipocyte biology via adjacent
binding on a genome-wide scale. Genes Dev. 22, 2941–2952.
Li, B., and Dewey, C.N. (2011). RSEM: accurate transcript quantification from
RNA-Seq data with or without a reference genome. BMC Bioinformatics 12,
323.
Li, H., and Durbin, R. (2009). Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25, 1754–1760.
Li, Z., Schug, J., Tuteja, G., White, P., and Kaestner, K.H. (2011). The nucleosome map of the mammalian liver. Nat. Struct. Mol. Biol. 18, 742–746.
Liu, L., Cheung, T.H., Charville, G.W., Hurgo, B.M., Leavitt, T., Shih, J., Brunet,
A., and Rando, T.A. (2013). Chromatin modifications as determinants of muscle stem cell quiescence and chronological aging. Cell Rep 4, 189–204.
Lund, E., Oldenburg, A.R., Delbarre, E., Freberg, C.T., Duband-Goulet, I., Eskeland, R., Buendia, B., and Collas, P. (2013). Lamin A/C-promoter interactions specify chromatin state-dependent transcription outcomes. Genome
Res. 23, 1580–1589.
Matys, V., Kel-Margoulis, O.V., Fricke, E., Liebich, I., Land, S., Barre-Dirrie, A.,
Reuter, I., Chekmenev, D., Krull, M., Hornischer, K., et al. (2006). TRANSFAC
and its module TRANSCompel: transcriptional gene regulation in eukaryotes.
Nucleic Acids Res. 34 (Database issue), D108–D110.
McLean, C.Y., Bristor, D., Hiller, M., Clarke, S.L., Schaar, B.T., Lowe, C.B.,
Wenger, A.M., and Bejerano, G. (2010). GREAT improves functional interpretation ofcis-regulatory regions. Nat. Biotechnol. 28, 495–501.
Niedernhofer, L.J., Garinis, G.A., Raams, A., Lalai, A.S., Robinson, A.R., Appeldoorn, E., Odijk, H., Oostendorp, R., Ahmad, A., van Leeuwen, W., et al.
(2006). A new progeroid syndrome reveals that genotoxic stress suppresses
the somatotroph axis. Nature 444, 1038–1043.
Panasyuk, G., Espeillac, C., Chauvin, C., Pradelli, L.A., Horie, Y., Suzuki, A.,
Annicotte, J.S., Fajas, L., Foretz, M., Verdeguer, F., et al. (2012). PPARg contributes to PKM2 and HK2 expression in fatty liver. Nat. Commun. 3, 672.
Rakhshandehroo, M., Knoch, B., Müller, M., and Kersten, S. (2010). Peroxisome proliferator-activated receptor alpha target genes. PPAR Res. 2010,
2010.
Robinson, M.D., McCarthy, D.J., and Smyth, G.K. (2010). edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140.
Robinson, J.T., Thorvaldsdóttir, H., Winckler, W., Guttman, M., Lander, E.S.,
Getz, G., and Mesirov, J.P. (2011). Integrative genomics viewer. Nat. Biotechnol. 29, 24–26.
Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors 1005
Rodriguez, A., Muller, D.C., Metter, E.J., Maggio, M., Harman, S.M., Blackman, M.R., and Andres, R. (2007). Aging, androgens, and the metabolic
syndrome in a longitudinal study of aging. J. Clin. Endocrinol. Metab. 92,
3568–3572.
Swift, J., Ivanovska, I.L., Buxboim, A., Harada, T., Dingal, P.C., Pinter, J., Pajerowski, J.D., Spinler, K.R., Shin, J.W., Tewari, M., et al. (2013). Nuclear laminA scales with tissue stiffness and enhances matrix-directed differentiation.
Science 341, 1240104.
Rozowsky, J., Euskirchen, G., Auerbach, R.K., Zhang, Z.D., Gibson, T., Bjornson, R., Carriero, N., Snyder, M., and Gerstein, M.B. (2009). PeakSeq enables
systematic scoring of ChIP-seq experiments relative to controls. Nat. Biotechnol. 27, 66–75.
Talbert, P.B., and Henikoff, S. (2010). Histone variants—ancient wrap artists of
the epigenome. Nat. Rev. Mol. Cell Biol. 11, 264–275.
Savage, D.B. (2009). Mouse models of inherited lipodystrophy. Dis. Model.
Mech. 2, 554–562.
Tropberger, P., Pott, S., Keller, C., Kamieniarz-Gdula, K., Caron, M., Richter,
F., Li, G., Mittler, G., Liu, E.T., Bühler, M., et al. (2013). Regulation of transcription through acetylation of H3K122 on the lateral surface of the histone octamer. Cell 152, 859–872.
Scaffidi, P., and Misteli, T. (2006). Lamin A-dependent nuclear defects in human aging. Science 312, 1059–1063.
Schumacher, B., Hoeijmakers, J.H., and Garinis, G.A. (2009). Sealing the gap
between nuclear DNA damage and longevity. Mol. Cell. Endocrinol. 299,
112–117.
Shackleton, S., Lloyd, D.J., Jackson, S.N., Evans, R., Niermeijer, M.F., Singh,
B.M., Schmidt, H., Brabant, G., Kumar, S., Durrington, P.N., et al. (2000).
LMNA, encoding lamin A/C, is mutated in partial lipodystrophy. Nat. Genet.
24, 153–156.
Sun, K., Battle, M.A., Misra, R.P., and Duncan, S.A. (2009). Hepatocyte
expression of serum response factor is essential for liver function, hepatocyte
proliferation and survival, and postnatal body growth in mice. Hepatology 49,
1645–1654.
Sun, Z., Miller, R.A., Patel, R.T., Chen, J., Dhir, R., Wang, H., Zhang, D., Graham, M.J., Unterman, T.G., Shulman, G.I., et al. (2012). Hepatic Hdac3 promotes gluconeogenesis by repressing lipid synthesis and sequestration.
Nat. Med. 18, 934–942.
Trapnell, C., Pachter, L., and Salzberg, S.L. (2009). TopHat: discovering splice
junctions with RNA-Seq. Bioinformatics 25, 1105–1111.
Umlauf, D., Goto, Y., and Feil, R. (2004). Site-specific analysis of histone
methylation and acetylation. Methods Mol. Biol. 287, 99–120.
Vlieghe, D., Sandelin, A., De Bleser, P.J., Vleminckx, K., Wasserman, W.W.,
van Roy, F., and Lenhard, B. (2006). A new generation of JASPAR, the
open-access repository for transcription factor binding site profiles. Nucleic
Acids Res. 34 (Database issue), D95–D97.
Willis-Martinez, D., Richards, H.W., Timchenko, N.A., and Medrano, E.E.
(2010). Role of HDAC1 in senescence, aging, and cancer. Exp. Gerontol. 45,
279–285.
Workman, J.L., and Kingston, R.E. (1992). Nucleosome core displacement
in vitro via a metastable transcription factor-nucleosome complex. Science
258, 1780–1784.
Zambelli, F., Pesole, G., and Pavesi, G. (2013). PscanChIP: Finding over-represented transcription factor-binding site motifs and their correlations in sequences from ChIP-Seq experiments. Nucleic Acids Res. 41 (Web Server
issue), W535–43.
1006 Cell Reports 9, 996–1006, November 6, 2014 ª2014 The Authors